Observational Methods

Research methods in psychology

This post was published in Psychology Review

All sciences are based on observation, in the broadest sense of the term. In this sense, ‘observation’ is not the same as ‘looking at things’. It refers to any means of gathering data about the world around us, including by means of specialist equipment that enables us to ‘observe’ the furthest reaches of the universe or to ‘observe’ the behaviour of neurons in the human brain.

All research in psychology is, by definition, based on observation. We observe the answers that our participants give on our questionnaires, the responses of interviewees to our semi-structured interviews, and the performances of participants in our experiments (the dependent variable is often measured by observation). But when we talk about ‘observational methods’ we are really referring to a set of methods that are, more literally, based on watching (and listening to) things that happen, and recording stuff about those things.

Broadly speaking, and purely for the purposes of explanation, it can be helpful to divide the world of observational methods into two camps: unstructured and structured.

Unstructured observation

When we talk of unstructured observational methods we are generally referring to research methods that involve observing a social setting, an institution, an organisation, a ‘case’, or such like, with a view to writing a richly detailed account of the observed thing. Participant observation is one particular tradition of inquiry that tends to draw on unstructured observations. In participant observation, researchers aim to become members of a social group, within a particular setting, with a view to observing that group, holistically, ‘from within’.

For example, the famous study by Rosenhan (1973) saw researchers getting themselves admitted to psychiatric hospitals as patients, so that they could covertly observe what was going on as members of the institution that they had set out to study. Such methods are usually associated with the collection and analysis of qualitative data. They tend to result in engaging and insightful narrative accounts of the thing that is being studied, albeit the problem that such methods do not easily lend themselves to replication.

Participant observation tends to be ethically challenging. By its very nature, it tends to be done ‘covertly’, such that the people being observed do not know that they are being observed. Covert observation can only be justified in certain circumstances and is more often associated with undercover journalism, than with academic psychology, because the benefits of the observation must demonstrably outweigh the ethical costs.

Covert and overt observations 

Covert observation takes place when people being observed are not aware they are being observed. In an overt observation, by contrast, there is no particular effort made to conceal the observation process. Many of the classic participant observations, such as that by Rosenhan, are covert. However, both participant and non-participant observations can be overt or covert.

One clear example of the benefits of participant observations is the 2016 Panorama investigation of neglect and abuse at a Cornish care home. Because of the way people ‘react’ to being observed, it is difficult to see how this kind of observation could be done, effectively, in any other way. Without special justification, observing those who do not know they are being observed should be limited to situations in which people are ‘in public’ and subject to the everyday observations of others (see the code of conduct and ethical guidelines o of the British Psychological Society, 2009).

Structured observation

Structured observational methods, by way of contrast, take a more ‘reductive’ (i.e. simplified) approach to data collection. Structured methods are based on observational schedules—checklists—which determine what is (and, crucially, what is not) to be recorded. Structured observations usually involve collecting quantitative data which tell us about the frequencies, durations, and/or sequences of the things that are being observed.

For example, Bakeman and Brownlee (1980) observed the sequences of different types of play (‘unoccupied’, ‘solitary’, ‘together’, ‘parallel’, ‘group’) in which 3-year-old children engaged. They were able to conclude that parallel play (children playing alongside each other) acts as a ‘bridge’ to group play (children playing with each other).

Structured observational methods place great store on the importance of replication. Specifically, a great deal of effort in any structured observational study goes into demonstrating that independent observers can use the same observational schedule, to observe the same sequences of behaviour, and obtain the same findings. This is known as the process of establishing ‘inter-observer reliability’.

Process and outcome

One distinctive feature of observational methods is that they are well placed to help us understand ‘process’, in contrast to much psychological research which tends to focus on ‘outcome’. Take, for example, communication interventions for the challenges faced by children on the autism spectrum, children with Down syndrome, and children with developmental delay. Outcome studies show how effective an intervention might (or might not) be. But systematic observation studies like Brown and Woods (2016) can show how the intervention works, and provide more detail about which aspects of the process are associated with the outcomes. For example, Brown and Woods were able to show that parents were more likely to use communication strategies that they had learned immediately after certain kinds of coaching session, and that the children’s use of the skills they had learned was associated with certain types of parental interactions. Examining the process of the intervention, in this way, is like identifying its ‘active ingredients’ (Brown and Woods, 2016, p. 121).

Observer effects

One key question to ask in relation to all types of observation is, ‘what effect is the observation having on the thing that is being observed?’ This is a difficult concept to get hold of in relation to observations of the physical world where quantum physics tells us that the more sure we are about where something is, the less sure we are about its velocity and vice versa. This observer effect is, however, relatively straightforward to understand in human terms. People ‘react’ (hence the effect is referred to as ‘reactivity’) to being observed and behave accordingly. As soon as you point a camera at someone to take a picture, they change their behaviour and put on their ‘picture face’. Just observing them changes the moment, so our pictures don’t produce a record of our lives but instead a series of posed images. So, when we observe people we might be recording data about their behaviour when being observed, rather than data about their behaviour in everyday life. And those might be two rather different things.

And then there is the issue of observer bias. There is no getting away from the issue that we tend to see what we want to see and hear what we want to hear. In our observations we focus on the things that are most important or most interesting to us and we are more likely to record these and even distort the record. Just ask fans of opposing football teams to describe the same match and you’ll see the issue.

References

Bakeman, R. and Brownlee, J. R. (1980) The strategic use of parallel play: A sequential analysis. Child Development, Vol. 51, No. 3, pp. 873-878.

BBC Panorama. (2016) Nursing homes undercover. London: BBC.

Brown, J. A. and Woods, J. J. (2016) Parent-implemented communication intervention. Topics in Early Childhood Special Education, Vol. 36, No. 2, pp. 115-124.

Ethics Committee of the British Psychological Society (2009) Code of ethics and conduct: Guidance published by the ethics committee of the British Psychological Society. Leicester: The British Psychological Society.

Rosenhan, D. L. (1973) On being sane in insane places. Science, Vol. 179, No. 4070, pp. 250-258.

Published originally as:

Grayson, A. (2018). Observational methods. Psychology Review, 24(2), 9–11.